基于双目视觉的水面障碍物识别算法

Recognition Algorithm of Water Surface Obstacle Based on Binocular Vision

为了实现无人艇对水面障碍物的定位,将双目视觉技术应用于障碍物图像识别与定位,提出一种改进的双目识别的快速算法。该算法基于HSI颜色空间和SURF特征原理实现特征点的提取,运用多类约束条件对匹配点对进行筛选,提高匹配效率。

In order to realize the positioning of unobstructed objects on water surface obstacles, binocular vision technology is applied to obstacle image recognition and positioning, and an improved algorithm for binocular recognition is proposed. The algorithm extracts feature points based on HSI color space and SURF feature principle. It uses multiple constraints to screen matching pairs and improves matching efficiency.